@machesterdragon Hello. Yes, it can never have too many examples 🙂

Inside the documentation sections, there are examples of strategies and technical indicators.
It's better to keep track of new strategy examples or updates on GitHub: https://github.com/quantiacs

@machesterdragon said in Cannot cast ufunc 'multiply' output from dtype('float64') to dtype('int64') with casting rule 'same_kind':

weights *= asset_filter

Hello. Change

weights *= asset_filter

to

weights = weights * asset_filter

You can see usage examples here:
https://quantiacs.com/documentation/en/user_guide/dynamic_assets_selection.html

Everything is working correctly now, try cloning the strategy.

@vg2001 Dear vg2001,
There were some problems with the servers, thank you for your patience.
Regards

@support Thank you !

@commanderangle Hi, the reason is that the algorithm is generating non-negative weights for assets which are not liquid at some point in time, in other words they are not part of the Nasdaq index at that time.

@multi_byte-wildebeest is it taking 2-3 minutes for each point in time?

@dark-pidgeot Hi, the question is a bit too generic, what do you need exactly to speed up?

@vyacheslav_b Thank you so much.

I have one more question for you to answer. I ran the precheck and the result was nan value the first time, but I set the min_date to 2005 - 01 - 01. I would like to ask, why is there a nan value problem? Is it because the ticker I chose had some companies that weren't listed at that time? My strategy id code is # 16767242. Thank you so much

Screenshot 2024-04-09 173002.png
Screenshot 2024-04-09 173012.png

@vg2001 of course you are right, thanks for pointing out the typo in the formula!

I finally resolved the issue, after lots of struggle. The custom layers, custom loss function and the function had to be serialized and deserialized correctly in order to save the architecture and weights as Json, rather than in a dictionary, like is suggested for pytorch in the Neural netowork template.
It seems Pytorch is way more user friendly when it comes to saving and loading models.

@vyacheslav_b Thank you very much for your support.

I would like to ask, if I want to filter out the crypto codes with the highest sharpness, what should I do? Thank you. I tried using the get_best_instruments function but it didn't work

import qnt.stats as qnstats # data = qndata.stocks.load_ndx_data(tail = 17*365, dims = ("time", "field", "asset")) data = qndata.stocks.load_ndx_data(min_date="2005-01-01") def get_best_instruments(data, weights, top_size): # compute statistics: stats_per_asset = qnstats.calc_stat(data, weights, per_asset=True) # calculate ranks of assets by "sharpe_ratio": ranks = (-stats_per_asset.sel(field="sharpe_ratio")).rank("asset") # select top assets by rank "top_period" days ago: top_period = 1 rank = ranks.isel(time=-top_period) top = rank.where(rank <= top_size).dropna("asset").asset # select top stats: top_stats = stats_per_asset.sel(asset=top.values) # print results: print("SR tail of the top assets:") display(top_stats.sel(field="sharpe_ratio").to_pandas().tail()) print("avg SR = ", top_stats[-top_period:].sel(field="sharpe_ratio").mean("asset")[-1].item()) display(top_stats) return top_stats.coords["asset"].values get_best_instruments(data, weights, 10)

19ae499c-71f3-4702-bba3-81d20fb6c5ac-image.png

@multi_byte-wildebeest Hello.

How to get the Sharpe Ratio is in the Quick Start template.
https://github.com/quantiacs/strategy-q20-nasdaq100-quick-start/blob/master/strategy.ipynb

import qnt.stats as qnstats def get_sharpe(market_data, weights): rr = qnstats.calc_relative_return(market_data, weights) sharpe = qnstats.calc_sharpe_ratio_annualized(rr).values[-1] return sharpe sharpe = get_sharpe(data, weights) # weights.sel(time=slice("2006-01-01",None))

or

import qnt.output as qnout qnout.check(weights, data, "stocks_nasdaq100")

or

stat = qnstats.calc_stat(data, weights) display(stat.to_pandas().tail())

or

import qnt.graph as qngraph statistics = qnstats.calc_stat(data, weights) display(statistics.to_pandas().tail()) performance = statistics.to_pandas()["equity"] qngraph.make_plot_filled(performance.index, performance, name="PnL (Equity)", type="log") display(statistics[-1:].sel(field=["sharpe_ratio"]).transpose().to_pandas()) qnstats.print_correlation(weights, data)

Please look at this post
https://quantiacs.com/community/topic/515/what-is-forward-looking-and-why-it-s-effective-badly-to-strategy/6?_=1711712434795

@vyacheslav_b perfect. It wasnt obvious to me that single pass was meant by that. Thank you

@illustrious-felice Hello. please look at this post
https://quantiacs.com/community/topic/515/what-is-forward-looking-and-why-it-s-effective-badly-to-strategy/6?_=1711712434795

@captain-dog Hi,

This is the total number of different companies that have been constituents of Nasdaq-100 index at some point, from 2006-01-01 and so on. At any point, there are 100 companies in index (sometimes more, up to 108) and those are considered as liquid ("is_liquid"=1.0), as long as they are part of the index.

@multi_byte-wildebeest Dear multi_byte-wildebeest,
You can check logs for your submission in the Filtered panel, so you can get more information why it might have failed. Please note that installing libraries takes time and should be placed in init.ipynb file and not in strategy.ipynb. Also check for potential bugs, errors and tricky loops.

Regards

@gjhernandezp Dear gjhernandezp,

No, models are trained from scratch taking the train_period of the data that is changing depending on the processing day. No, the retraining is done from scratch, taking the whole available data up until that point in time. That is not available at the moment, but take a note that when submitting a ML strategy it should be in a single-pass mode. For example instead of using backtest_ml function that is designed for testing purposes, one should use single-pass, for example: data_train = load_data(train_period) models = train_model(data_train) data_predict = load_data(lookback_period) weights_predict = predict(models, data_predict) print_stats(data_predict, weights_predict) qnout.write(weights_predict) # To participate in the competition, save this code in a separate cell.

@multi_byte-wildebeest Hello. I don't use machine learning models in trading.

@support Hello,

got it, thanks for the reply,

@illustrious-felice
Hello, sorry for late reply, can you please let us know about the progressbar2 version installed from environment you are currently using? Be sure that "progressbar2" package is installed and not "progressbar".

Best regards,